This project explored the reliability and utility of transcription in coding qualitative data across two studies in a program evaluation context. The first study tested the method of direct audio coding, or coding audio files without transcripts, using qualitative data software. The presence and frequency of codes applied in direct audio coding and traditional transcription coding were compared and the two methods produced similar results. Direct audio coding was then employed in an evaluation study to monitor implementation and the method was found to be reliable. Implications are discussed with considerations for both researchers and practitioners.